Systems and methods for video processing based on motion-aligned spatio-temporal steering kernel regression
AssigneeSharp Laboratories of America, Inc.
InventorsPetrus J. L. Van Beek, Hiroyuki Takeda
Life of this patent
Estimating the term…
Abstract
A method for video processing based on motion-aligned spatio-temporal steering kernel regression may include estimating local spatio-temporal steering parameters at pixels in input video data. The method may also include using the local spatio-temporal steering parameters to apply motion-aligned spatio-temporal steering kernel regression to the input video data to perform at least one of upscaling and noise smoothing.
Claims18
A method for video processing based on motion-aligned spatio-temporal steering kernel regression, comprising:
- estimating local spatio-temporal steering parameters at pixels in input video data, wherein estimating the local spatio-temporal steering parameters comprises estimating at least one spatial steering parameter and estimating at least one motion steering parameter;
- using the local spatio-temporal steering parameters to apply motion-aligned spatio-temporal steering kernel regression to the input video data to perform at least one of upscaling and noise suppression;
- estimating spatial derivatives, wherein spatial steering parameter estimation and motion steering parameter estimation both depend on the spatial derivatives;
- and
- estimating temporal derivatives, wherein the motion steering parameter estimation depends on the temporal derivatives.
A computer system for video processing based on motion-aligned spatio-temporal steering kernel regression, comprising:
- a processor;
- memory in electronic communication with the processor;
- and
- instructions stored in the memory, the instructions being executable to:
- estimate local spatio-temporal steering parameters at pixels in input video data, wherein estimating the local spatio-temporal steering parameters comprises estimating at least one spatial steering parameter and estimating at least one motion steering parameter;
- use the local spatio-temporal steering parameters to apply motion-aligned spatio-temporal steering kernel regression to the input video data to perform at least one of upscaling and noise suppression;
- estimating spatial derivatives, wherein spatial steering parameter estimation and motion steering parameter estimation both depend on the spatial derivatives;
- and
- estimating temporal derivatives, wherein the motion steering parameter estimation depends on the temporal derivatives.
A non-transitory computer-readable medium comprising executable instructions for:
- estimating local spatio-temporal steering parameters at pixels in input video data, wherein estimating the local spatio-temporal steering parameters comprises estimating at least one spatial steering parameter and estimating at least one motion steering parameter;
- using the local spatio-temporal steering parameters to apply motion-aligned spatio-temporal steering kernel regression to the input video data to perform at least one of upscaling and noise suppression;
- estimating spatial derivatives, wherein spatial steering parameter estimation and motion steering parameter estimation both depend on the spatial derivatives;
- and
- estimating temporal derivatives, wherein the motion steering parameter estimation depends on the temporal derivatives.
Dependencies are inferred from the claim text.